Roadway, traffic volume and crash data were analyzed to identify sites that are good candidates for improvement for weatherrelated crashes. These sites with promise were then shared with meteorology researchers on the team who used them to explore and identify significant weather-related signatures within meteorological data. These analyses were undertaken to determine the feasibility of an Advanced Road Safety and Weather Warning System, which would provide forecasts of significant weather events that could be broadcast to the public via communications outlets such as highway advisory radio, websites, changeable message signs and direct media broadcasts. The linked analysis of crash and meteorological data was a success. While the specific findings of this particular study are applicable to PennDOT District 2-0, the methodology is applicable to any other PennDOT district with comparable data and extendable to any type of crash under investigation and even to other facility types. This approach can be particularly helpful when analyzing crash types that are relatively infrequent; the use of random effects might provide a means for accounting for random variability. (Author/publisher)
Abstract